Nitric acid uptake on subtropical cirrus cloud particles

P. J. Popp,1,2 R. S. Gao,3 T. P. Marcy,1,2 D. W. Fahey,1,2 P. K. Hudson,1,2 T. L. Thompson,1

B. Karcher,4 B. A. Ridley,5 A. J. Weinheimer,5 D. J. Knapp,5 D. D. Montzka,5

D. Baumgardner,6 T. J. Garrett,7 E. M. Weinstock,8 J. B. Smith,8 D. S. Sayres,8

J. V. Pittman,8 S. Dhaniyala,9,10 T. P. Bui,11 and M. J. Mahoney12

Received 16 October 2003; revised 9 January 2004; accepted 2 February 2004; published 17 March 2004.

[1] The redistribution of HNO3 via uptake and sedimentation by cirrus cloud particles isconsidered an important term in the upper tropospheric budget of reactive nitrogen.Numerous cirrus cloud encounters by the NASA WB-57F high-altitude research aircraftduring the Cirrus Regional Study of Tropical Anvils and Cirrus Layers-Florida AreaCirrus Experiment (CRYSTAL-FACE) were accompanied by the observation ofcondensed-phase HNO3 with the NOAA chemical ionization mass spectrometer. Theinstrument measures HNO3 with two independent channels of detection connected toseparate forward and downward facing inlets that allow a determination of the amount ofHNO3 condensed on ice particles. Subtropical cirrus clouds, as indicated by the presenceof ice particles, were observed coincident with condensed-phase HNO3 at temperaturesof 197–224 K and pressures of 122–224 hPa. Maximum levels of condensed-phaseHNO3 approached the gas-phase equivalent of 0.8 ppbv. Ice particle surface coveragesas high as 1.4 � 1014 molecules cm�2 were observed. A dissociative Langmuir adsorptionmodel, when using an empirically derived HNO3 adsorption enthalpy of�11.0 kcal mol�1,effectively describes the observed molecular coverages to within a factor of 5. Thepercentage of total HNO3 in the condensed phase ranged from near zero to 100% in theobserved cirrus clouds. With volume-weighted mean particle diameters up to 700 mmand particle fall velocities up to 10 m s�1, some observed clouds have significant potentialto redistribute HNO3 in the upper troposphere. INDEX TERMS: 0305 Atmospheric Composition

and Structure: Aerosols and particles (0345, 4801); 0320 Atmospheric Composition and Structure: Cloud

physics and chemistry; 0322 Atmospheric Composition and Structure: Constituent sources and sinks;

KEYWORDS: cirrus clouds, nitric acid, uptake, ice particles

Citation: Popp, P. J., et al. (2004), Nitric acid uptake on subtropical cirrus cloud particles, J. Geophys. Res., 109, D06302,

doi:10.1029/2003JD004255.

1. Introduction

[2] Cirrus clouds are ubiquitous throughout the uppertroposphere (UT) and can cover as much as 40% of Earth’ssurface [Liao et al., 1995; Jin et al., 1996; Wang et al.,1996; Wylie and Menzel, 1999]. Composed of ice crystals[Lynch, 2002], cirrus clouds are known to play a complexand significant role in the global radiation budget [Liou,1986]. Cirrus clouds can be formed in situ in the UT[Karcher, 2002], as a result of synoptic weather disturban-ces, or in the anvil outflow at the top of cumulonimbusclouds [Sassen, 2002]. Tropical cirrus clouds around thepeak convective detrainment level are formed primarily viathe latter mechanism, and can reach altitudes of up to 18 kmwhen produced in deep convective systems. The broadlateral and vertical extent of anvil cirrus clouds producedin the tropics is expected to exert a greater influence onEarth’s climate system than midlatitude cirrus [Heymsfieldand McFarquhar, 2002]. Owing to the high altitudes andoften-remote locations of tropical cirrus, however, compre-hensive in situ measurements of these clouds have beenlimited.

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, D06302, doi:10.1029/2003JD004255, 2004

1Cooperative Institute for Research in Environmental Sciences,University of Colorado, Boulder, Colorado, USA.

2Also at Aeronomy Laboratory, National Oceanic and AtmosphericAdministration, Boulder, Colorado, USA.

3Aeronomy Laboratory, National Oceanic and Atmospheric Adminis-tration, Boulder, Colorado, USA.

4Institut fur Physik der Atmosphare, Deutsches Zentrum fur Luft- andRaumfahrt Institut fur Physik der Atmosphare, Wessling, Germany.

5Atmospheric Chemistry Division, National Center for AtmosphericResearch, Boulder, Colorado, USA.

6Universidad Nacional Autonoma de Mexico, Centro de Ciencias de laAtmosfera, Ciudad Universitaria, Mexico City, Mexico.

7Department of Meteorology, University of Utah, Salt Lake City, Utah,USA.

8Atmospheric Research Project, Harvard University, Cambridge,Massachusetts, USA.

9Division of Geology and Planetary Sciences, California Institute ofTechnology, Pasadena, California, USA.

10Now at Department of Mechanical and Aeronautical Engineering,Clarkson University, Potsdam, New York, USA.

11NASA Ames Research Center, Moffett Field, California, USA.12Jet Propulsion Laboratory, California Institute of Technology,

Pasadena, California, USA.

Copyright 2004 by the American Geophysical Union.0148-0227/04/2003JD004255$09.00

D06302 1 of 14

[3] Model simulations by Lawrence and Crutzen [1998]suggest that the uptake and gravitational redistribution ofnitric acid (HNO3) by cirrus cloud particles may represent asignificant sink of HNO3 in the UT. Nitric acid serves asa primary reservoir species for nitrogen oxides (NOx)[Neuman et al., 2001], which are directly involved in thephotochemical production of tropospheric ozone [Jaegle etal., 1998]. Model studies of cirrus-processed air havedemonstrated that the sedimentary removal of HNO3 fromthe UT can effect strong local reductions in NOx, with theconsequence of significant reductions in the net ozoneproduction rate [Meier and Hendricks, 2002]. Since ozoneis known to be an effective greenhouse gas in the tropo-sphere [Albritton et al., 2001], particularly near the tropo-pause [Lacis et al., 1990], understanding the uptake andredistribution of HNO3 by cirrus cloud particles may beimportant in assessing the contribution of cirrus clouds tothe radiative forcing of climate change.[4] A number of laboratory studies have investigated the

uptake of HNO3 on ice surfaces at temperatures typical ofthe UT. Experiments performed by Zondlo et al. [1997] onvapor-deposited ice films at 211 K resulted in observedHNO3 surface coverages of 1.5 � 1015 molecules cm�2. Aseries of similar experiments reported by Hudson et al.[2002] at an HNO3 pressure (P(HNO3)) of 1.1 � 10�6 hPaindicated a negative temperature dependence to the observeduptake, with coverages of 1.1� 1014 to 5.9� 1013 moleculescm�2 over a temperature range of 214–220 K. Uptakestudies performed on ice films by Abbatt [1997] yieldedcoverages of up to 2.9 � 1014 molecules cm�2 at temper-atures as low as 208 K, with no apparent dependency onP(HNO3) values over the range 1.7 � 10�7 to 4.1 �10�6 hPa. Hynes et al. [2002] reported comparable cover-ages using a similar technique, although they observedcoverages increasing by factor of 2 over a nearly 10-foldincrease in P(HNO3), from 5.0 � 10�7 to 3.0 � 10�6 hPa. Alaboratory study of HNO3 uptake on nebulized half-microndiameter ice particles at 230 K yielded coverages similar tothose observed on the ice films (1.2 � 1014 moleculescm�2), although these experiments were performed at therelatively high P(HNO3) of 7 � 10�6 hPa [Arora et al.,1999]. There has not yet been an extensive laboratory studyof HNO3 uptake on ice surfaces performed at P(HNO3)values typical of the UT (<2.0 � 10�7 hPa).[5] Prior field studies of HNO3 uptake on cirrus cloud

particles have been made at mid and high latitudes. Mea-surements of total reactive nitrogen (NOy = NO + NO2 +2N2O5 + HNO3 +. . .) in a mountain wave cloud over thecontinental United States reported by Weinheimer et al.[1998] indicated that levels of condensed NOy in the cloudapproached 20% of total NOy. Surface coverages on thewave cloud ice particles were calculated to be as high as2.5 � 1013 molecules cm�2 [Hudson et al., 2002]. Measure-ments of condensed-phase NOy in cirrus layers in the ArcticUT by Kondo et al. [2003] yielded HNO3 coverages as highas 1.6 � 1014 molecules cm�2 at temperatures of approx-imately 200 K, with coverages decreasing at warmer tem-peratures. Meilinger et al. [1999] conducted similarmeasurements in Arctic cirrus clouds at 196 K and reportedcoverages of only 1 � 1013 molecules cm�2. An extensivedataset of measurements made at midlatitudes revealedmedian levels of condensed NOy (assumed to be HNO3)

in the Northern Hemisphere of 3.6 � 1012 molecules cm�2

to be greater than twice that observed in the SouthernHemisphere [Ziereis et al., 2004].[6] We report here an extensive dataset of in situ mea-

surements, including gas- and condensed-phase HNO3 andice particle surface area density (SAD), obtained in sub-tropical in situ and anvil cirrus clouds. These measurementswere conducted onboard the NASA WB-57F high-altituderesearch aircraft as part of the Cirrus Regional Study ofTropical Anvils and Cirrus Layers Florida Area CirrusExperiment (CRYSTAL-FACE) mission. The data are usedhere to assess the uptake of HNO3 by subtropical cirruscloud particles and explore the partitioning of HNO3 be-tween the gas and ice particle phases in cirrus clouds.

2. Instrumentation

[7] This study utilizes data from a number of in situinstruments onboard the NASAWB-57F aircraft. Gas-phaseand condensed-phase HNO3 measurements made by chem-ical ionization mass spectrometry are described in moredetail below. Particle size distribution and number densitymeasurements made by the Cloud, Aerosol and Precipita-tion Spectrometer (CAPS) were used to derive SAD andvolume-weighted mean diameter (VMD) for ice particles inthe size range between 0.35–1550 mm [Baumgardner et al.,2001]. Note that all particle sizes cited herein refer toparticle diameter, and not radius. A second, independentmeasurement of SAD was provided by the Cloud Integrat-ing Nephelometer (CIN) [Gerber et al., 2000]. Ice watercontent (IWC) and water (H2O) vapor were measured by theHarvard University Lyman-a hygrometer [Weinstock et al.,1994; E. M. Weinstock et al., manuscript in preparation,2004]. Nitric oxide (NO) and total reactive nitrogen (NOy)were measured by catalytic reduction and chemilumines-cence (A. J. Weinheimer et al., manuscript in preparation,2004). Ambient temperature and pressure, and WB-57F trueair speed were measured by the Meteorological Measure-ment System (MMS) [Scott et al., 1990]. The precision andaccuracy of these measurements are summarized in Table 1.Tropopause height was measured by the microwave tem-perature profiler (MTP) [Denning et al., 1989].[8] HNO3 was measured using the NOAA chemical

ionization mass spectrometer (CIMS) located in the thirdpallet position of the NASA WB-57F aircraft. This instru-ment measures HNO3 with an accuracy of ±20% andprecision of 30 pptv (1s, 10-s averages), and has beendescribed in detail elsewhere [Neuman et al., 2000]. Prior toCRYSTAL-FACE, the NOAA CIMS was modified by theaddition of a second independent channel for the measure-ment of HNO3 and the relocation of the original sampleinlet on the CIMS inlet pylon (Figure 1). The two CIMSchannels are designed to provide identical measurements ofgas-phase HNO3. Owing to differences in the particlesampling efficiencies of the two inlets, however, the twochannels have different sensitivity to condensed-phaseHNO3. When sampling in cirrus clouds, the forward facingfront inlet samples both gas-phase HNO3 and any HNO3

condensed on the cirrus particles. The downward facingbottom inlet samples primarily gas-phase HNO3 because theplane of the sampling orifice is parallel to the direction offlow over the inlet, which is set by the flow straightener.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

2 of 14

D06302

Semiempirical calculations indicate that approximately 50%of 0.1 mm particles and greater than 90% of 1 mm and largerparticles are inertially stripped from the air sampled by thebottom inlet [Vincent et al., 1986]. Calculations furthersuggest that ice particles greater than 10 mm in diameter(typical of cirrus clouds observed during CRYSTAL-FACE)are almost entirely removed from the sampled air. Thus, formost of the cirrus clouds sampled, the HNO3 from the

bottom inlet is taken to be a measure of the gas-phase HNO3

abundance.[9] The conclusion that ice particles greater than approx-

imately 1 mm are inertially separated from air sampled bythe bottom inlet is further supported by measurements madein the contrail of the WB-57F during CRYSTAL-FACE.The contrail contained ice particles with high numberdensities (100–200 cm�3) and volume-weighted mean

Table 1. Measurement Details

Measurement Institution Precisiona Accuracy Reference

HNO3 NOAA Aeronomy Lab 30 pptv, 10 pptvb ±20% Neuman et al. [2001]SAD-CAPS Univ. Nacional Autonoma de Mexico - ±50% Baumgardner et al. [2001]SAD-CIN University of Utah - ±15% Gerber et al. [2000]IWC Harvard University 0.7 ppmv ±17% E. M. Weinstock et al. (manuscript in preparation, 2004)Water vapor Harvard University 0.5 ppmv ±5% Weinstock et al. [1994]NO NCAR 5 pptv ±6% A. J. Weinheimer et al. (manuscript in preparation, 2004)NOy NCAR 8 pptv ±12% A. J. Weinheimer et al. (manuscript in preparation, 2004)Temperature NASA Ames 0.01 K ±0.3 K Scott et al. [1990]Pressure NASA Ames 0.1 hPa ±0.3 hPa Scott et al. [1990]True air speed NASA Ames 0.1 m s�1 ±1 m s�1 Scott et al. [1990]

aPrecision values are reported for 10-s averages.b10 pptv is the precision for condensed-phase HNO3.

Figure 1. Schematic diagram of the NOAA CIMS inlet pylon. Labeled components are as follows:(a) front channel inlet, (b) bottom channel inlet, (c) zero gas addition, (d) calibration gas addition,(e) reagent gas carrier addition, (f ) reagent gas addition, (g) HNO3 permeation cell (calibration standard),(h) flow control valve motor, (i) flow control valve body, ( j) flow tube, (k) ion source, and (l) flowstraightener. For clarity, complete components are shown only for the front channel. Inlet lines areconstructed of Teflon

TMtubing (6.4 mm outside diameter, 4.0 mm inside diameter), heated to temperatures

of 40�C (bottom channel inlet) and 48�C (front channel inlet) to avoid wall losses. The highertemperature of the front inlet ensures that ice particles entering the inlet evaporate upon impaction withthe tubing wall. The operating principles of the NOAA CIMS instrument have been described in detail byNeuman et al. [2000].

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

3 of 14

D06302

diameters on the order of 2 mm [Gao et al., 2003]. HNO3

was present in the contrail because of mixing between theexhaust gases and ambient air containing approximately0.4 ppbv HNO3. Measurements in the contrail as soon as4 min after formation indicated a difference between thefront and bottom CIMS channels. The minimum signal fromthe bottom channel was near zero inside the contrail. Thislow HNO3 signal is consistent with the removal of gas-phase HNO3 by uptake onto the 2-mm ice particles in thecontrail, and the inertial stripping of these particles from thebottom inlet sample flow. As expected, a simultaneousincrease in HNO3 above ambient values occurred in thefront CIMS channel, which does not discriminate against2 mm particles. If the bottom CIMS inlet sampled 2 mmparticles with any significant efficiency, HNO3 observed inthe bottom channel during the contrail intercept would notbe significantly lower than the ambient values immediatelyoutside the contrail.[10] The front CIMS inlet samples subisokinetically,

meaning the sample air velocity inside the inlet (U) is lessthan the WB-57F true air speed (U0) of 140–200 m s�1 atsampling altitudes. As a result, cirrus cloud particle numberdensities in the sampled air stream are inertially enhancedrelative to those in the ambient air. A computational fluiddynamics program (Fluent Inc., New Hampshire) was usedto estimate particle enhancement factors (EF) in the frontinlet by simulating the flow field and particle trajectories

around a two-dimensional horizontal cross section of theCIMS pylon and inlet structure (Figure 1). The value of EFis near unity for small particles (<0.1 mm) and increaseswith particle size, as found for similar configurations[Northway et al., 2002]. For particles larger than approxi-mately 10 mm in diameter, typical of cirrus cloud iceparticles sampled during CRYSTAL-FACE, EF for the frontinlet approaches the maximum value of U0/U. Since bothCIMS channels sample at a constant mass flow of 1.85standard liters per minute (slpm), U, and therefore, EF, aredependent upon the ambient temperature and pressure.Under typical WB-57F sampling conditions duringCRYSTAL-FACE (temperature = 213 K, pressure =170 hPa, U0 = 200 m s�1), EF has a maximum value ofapproximately 16.[11] Cirrus cloud particles entering the front CIMS inlet

travel through a 20 cm length of TeflonTMtubing (6.4 mm

outside diameter, 4.0 mm inside diameter) upstream of theCIMS flow control valve and flow tube (Figure 1). The useof Teflon

TMsample lines ensures that HNO3 will not readily

absorb on the inlet surfaces [Neuman et al., 1999]. Thistubing, which is heated to 48�C in flight, has two bends tohelp ensure that large particles entering the inlet will impacton the tubing walls and subsequently evaporate prior toreaching the flow control valve. Particles with diametersgreater than approximately 20 mm have large enoughstopping distances at the freestream velocity that they

Figure 2. Time series measurements of the HNO3 mixing ratio observed from the front and bottomCIMS channels (HNO3jfront and HNO3jgas, respectively,) on the flight of 11 July 2002. Note that values ofHNO3jfront do not include a correction for particle oversampling in the front channel inlet. Discontinuitiesin the time series result from CIMS instrument calibrations and other housekeeping procedures. Alsoshown are SAD, IWC, water vapor, relative humidity (with respect to ice), and ambient temperature andpressure. All data are represented as 10-s averages. Minor divisions on the horizontal scale represent15 min (or approximately 150 km) of flight.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

4 of 14

D06302

impact at the first bend. Some particles that do not fullyevaporate will impact in the body of the flow control valveor the flow tube entrance. HNO3 condensed on the particlesurfaces is liberated to the gas phase early in the evaporationprocess and measured as a gas-phase equivalent volumemixing ratio. The HNO3 mixing ratio measured by the frontCIMS channel, therefore, represents the sum of the gas- andthe condensed-phase values, with the condensed-phasecomponent enhanced by the value of EF.

3. Observations

[12] Condensed-phase HNO3 was observed coincidentwith cirrus cloud observations on 4 WB-57F science flightsconducted as part of CRYSTAL-FACE. These flights orig-inated and terminated at the United States Naval AirFacility, Key West (24.6�N, 81.7�W) in Florida on 11, 13,19, and 21 July 2002. Time series data of HNO3 mixingratios observed from the front and bottom CIMS channels(HNO3jfront and HNO3jgas, respectively), as well as particleSAD, IWC (represented as a gas-phase equivalent volumemixing ratio) and meteorological parameters are shown for11, 13, 19, and 21 July in Figures 2–5, respectively. Thepresence of cirrus cloud particles is indicated by increases inSAD and IWC above background values. The presence ofcondensed-phase HNO3 in a flight segment is indicated byHNO3jfront values that are significantly greater thanHNO3jgas values. Flight segments identified by purple bars

in panel (b) for 13 July (Figure 3) and 19 July (Figure 4)represent the observation of contrail cirrus clouds. As statedpreviously, these clouds are characterized by having highparticle number densities with volume-weighted meandiameters typically much lower than cirrus clouds formedby natural processes. Owing to the uncertainties in the SADmeasurements in the contrail-formed cirrus clouds and inthe value of EF for particles in this size range, these cloudsare not considered in the data analysis presented here.[13] Cirrus clouds were observed from the WB-57F at

pressures between 122 hPa and 224 hPa during the flightsof 11, 13, 19, and 21 July, corresponding to pressurealtitudes between 11 km and 15 km (Figures 2–5). Theseclouds were observed at temperatures between 197 K and224 K. Figures 2–5 show the strong temporal correlation of(HNO3jfront � HNO3jgas) with both SAD and IWC in cirrusclouds, with HNO3jgas approaching zero during a number ofcirrus events. Outside of clouds, measured values ofHNO3jfront and HNO3jgas generally agree well (with anoverall correlation coefficient, r, of 0.92). However, someperiods in Figures 2–5 show offsets between the twochannels that are best explained as changes in the inlet linesurfaces during the flight. Figure 4 also indicates elevatedvalues of relative humidity (with respect to ice) during anumber of cirrus cloud encounters, as described by Gao etal. [2003].[14] Condensed-phase HNO3, proportional to the differ-

ence between the values of HNO3jfront and HNO3jgas, was

Figure 3. Same as Figure 2, for the flight on 13 July 2002. Purple bars in Figure 3b represent 3 flightsegments in which cirrus clouds formed in the contrail of the WB-57F were observed. The inset inFigure 3a shows 1-s averages of HNO3jfront and HNO3jgas during the first contrail intercept. The purplebar in the inset represents the same time period as the first purple bar in Figure 3b. The vertical scale onthe inset panel is 0 to 1 ppbv. Measurements of HNO3jfront and HNO3jgas are incomplete during thesecond and third contrail intercepts due to instrument housekeeping procedures.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

5 of 14

D06302

Figure 4. Same as Figure 2, for the flight on 19 July 2002. Purple bars in Figure 4b represent 2 flightsegments in which cirrus clouds formed in the contrail of the WB-57F were observed. The insets inFigure 4a show 1-s averages of HNO3jfront and HNO3jgas during the two contrail intercepts. The purplebars in the first and second insets represent the same time periods as the first and second purple bars inFigure 4b. The vertical scale on the inset panels is 0–0.5 ppbv.

Figure 5. Same as Figure 2, for the flight on 21 July 2002.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

6 of 14

D06302

observed primarily at SADs greater than 200 mm2 cm�3

during CRYSTAL-FACE (Figure 6). Note that for thevalues of (HNO3jfront � HNO3jgas) shown in Figure 6,HNO3jfront is not corrected for particle oversampling. Valuesof (HNO3jfront�HNO3jgas) at SADs less than 200 mm2 cm�3

are near the detection limit and highly variable due to CIMSinstrument noise. In the analyses presented here, observa-tions at SADs greater than 200 mm2 cm�3 are selected torepresent measurements made in cirrus clouds (shown bythe dashed line in Figure 6).

3.1. Quantifying Condensed-Phase HNO3

[15] As stated previously, HNO3jfront represents the sum ofgas-phase and condensed-phase HNO3, with the condensed-phase component enhanced by the value of EF. The amountof HNO3 condensed on cirrus cloud particles (HNO3jcon) cantherefore be calculated according to equation (1),

HNO3jcon¼HNO3jfront � HNO3jgas

EF; ð1Þ

where HNO3jcon is reported as a gas-phase equivalentvolume mixing ratio with a precision of 10 pptv (1s, 10-saverages). The use of equation (1) in calculating HNO3jconis illustrated in Figure 7 for a cirrus cloud encounter by theWB-57F on 13 July 2002. Increases in SAD and IWC

during this cloud event are accompanied by an increase inHNO3jfront above the gas-phase value of approximately0.5 ppbv (Figures 7a, 7d, and 7e). Accounting for theparticle enhancement factor of approximately 13.7 usingequation (1), maximum values of HNO3jcon during thiscloud event approached 0.1 ppbv (Figures 7b and 7c). Itshould be noted here that calculated values of HNO3jcon incirrus clouds are not always consistent with the observeddecreases in HNO3jgas that result from HNO3 uptake.Quantitative agreement between HNO3jcon and deficits inHNO3jgas can only occur if the cloud particles are sampledin the same air mass in which uptake occurred. Owing togravitational settling, however, cirrus particles may sedi-ment into air masses that may be more or less depleted ingas-phase HNO3 at the time of sampling.

3.2. Cirrus Cloud Particle Measurements

[16] Cirrus cloud particle SAD was derived from mea-surements provided by both the CAPS and CIN instruments

Figure 6. (HNO3jfront � HNO3jgas) versus SAD for theflights of 11, 13, 19, and 21 July 2002. Values of(HNO3jfront � HNO3jgas) are proportional but not equal tocondensed-phase HNO3 (HNO3jcon) because they do notinclude a correction for particle oversampling in the frontchannel inlet. Red circles represent 10-s averages and blacksquares represent mean values of the 10-s data grouped intodeciles. Vertical bars represent the standard deviation aboutthe mean value in each decile, and horizontal bars indicatethe upper and lower boundaries of each decile. Note that thevertical axis is shown with a logarithmic scale at valuesgreater than 2.5. The dashed vertical line at the SAD of200 mm2 cm�3 represents the lower limit chosen here torepresent measurements made in cirrus clouds.

Figure 7. Calculation of HNO3jcon for a cirrus cloudencounter by the NASAWB-57F on 13 July 2002, showing(a) HNO3jfront and HNO3jgas, (b) EF, (c) HNO3jcon, (d) SAD,and (e) IWC. All data are shown as 10-s averages. Thehorizontal axis spans approximately 16 min and 180 km offlight.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

7 of 14

D06302

onboard the WB-57F during CRYSTAL-FACE. SAD wasderived from the CAPS data by integrating particle sizedistribution and number density measurements, while bulkmeasurements of cloud extinction coefficient at 635 nm wereused to derive SAD from the CIN data. A comparisonbetween the CIN- and CAPS-derived SADs indicates goodagreement between the two instruments for the flights on 11,19, and 21 July, with SADs derived from the CIN measure-ments ranging from 23–39% higher than the CAPS-derivedvalues on those 3 flight days (Figure 8). On 13 July the CINmeasurements were 54% higher than the CAPS values,which may be attributable to the sampling of optically thinsubvisual cirrus clouds near the tropopause on that day.Cloud extinction in these subvisual cirrus is close to thesensitivity threshold of the CIN. The observed differencesare nonetheless within the combined uncertainties of the twoinstruments on all 4 flight days. The analyses presented heremake use of SADs derived from the CAPS measurements.Owing to the fact that the SADs derived from the CINmeasurements are 23–54% higher than the CAPS values,the SADs utilized here can be considered lower limits.[17] As stated previously, the value of EF used in

calculating HNO3jcon approaches a maximum value of U0/U (the ratio of the WB-57F true air speed to the air velocityin the sample inlet) at particle sizes greater than approxi-mately 10 mm. SADs calculated from the CAPS measure-ments on the 4 flight days considered here (in cirrus cloudswith total SADs greater than 200 mm2 cm�3) indicate that92 ± 9% of the surface area resides on particles larger than10 mm in diameter. Use of the maximum value of EF incalculating HNO3jcon via equation (1), therefore, is expected

to introduce no more than 10% uncertainty into the value ofHNO3jcon.[18] The VMD of cirrus cloud particles observed during

CRYSTAL-FACE ranged from approximately 3 mm up to700 mm, with most clouds having VMDs greater than 20 mm(Figure 9a). The largest particles (>500 mm) were observedprimarily in clouds with SADs greater than 104 mm2 cm�3.As expected, IWC shows a strong correlation with SAD(Figure 9b). IWCs as high as 1000 ppmv were observedduring some cloud events (Figure 9b). The highest values ofVMD and IWC were observed from the WB-57F primarilyat temperatures between 205 K and 215 K during the flightsshown in Figure 9.

4. Discussion

4.1. HNO3 Uptake on Cirrus Cloud Particles

[19] The coincident observation of cirrus clouds andcondensed-phase HNO3 during CRYSTAL-FACE is as-sumed here to result from the uptake of HNO3 on thesurface of cirrus cloud particles. Laboratory studies indicatethat the low solubility of HNO3 in ice will not allow asignificant fraction of HNO3jcon to reside in the bulk of thecirrus particles [Sommerfeld et al., 1998; Hanson andRavishankara, 1991]. Furthermore, Domine and Thibert[1996] have suggested that the high diffusivity of HNO3 inice is such that HNO3 trapped in the bulk ice duringparticle formation will migrate to the particle surface. Wenote, however, that the measurements presented here can-not distinguish between surface uptake and HNO3 that maybe condensed in the bulk of the particles. HNO3 uptake oncirrus cloud particles can be represented in terms ofmolecular coverage, given by the ratio of HNO3jcon toSAD in units of molecules cm�2. HNO3 surface coveragesobserved during CRYSTAL-FACE are shown as a functionof temperature in Figure 10, with symbols colored accord-ing to P(HNO3) (see legend). Data shown by triangles attemperatures less than 200 K in Figure 10 representobservations under conditions in which nitric acid trihy-drate (NAT) is stable, as predicted by ambient temperatureand the ratio of HNO3jgas to H2O vapor [Hanson andMauersberger, 1988; Gao et al., 2003]. It has beenproposed that, under conditions in which NAT is stable,HNO3 forms NAT clusters or layers on the particle surfacewhich interfere with the condensation of H2O molecules onthe particle surface, and thereby increase the relativehumidity with respect to ice in the cirrus cloud [Gao etal., 2003].[20] The mean HNO3 coverage observed during

CRYSTAL-FACE was 1.9 � 1013 molecules cm�2, withmaximum coverages reaching 1.4 � 1014 molecules cm�2

during a few cirrus cloud events (Figure 10). While thegreatest coverages were observed at temperatures between205 K and 210 K, mean coverages binned according totemperature show no temperature dependence above 200 K(black symbols in Figure 10). The average value formeasurements between 195 K and 200 K is approximatelya factor of 3 greater than values above 200 K. Generallyhigher HNO3 coverages at lower temperatures have beenobserved in field measurements reported by both Kondo etal. [2003] and Ziereis et al. [2004]. HNO3 coverages show aminimal dependence on P(HNO3), with the lowest cover-

Figure 8. Comparison of CIN- and CAPS-derived SADmeasured during the flights of 11, 13, 19, and 21 July 2002.Data at SADs less than 3 mm2 cm�3 are expanded in theinset. All data are shown as 10-s averages. Lines represent aleast squares fit to the data for each flight date. The slope ofthe linear fit (constrained through the origin) is shown foreach flight date in the figure legend.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

8 of 14

D06302

ages occurring at P(HNO3) values below 2.5 � 10�8 hPa(Figure 10). A number of laboratory studies have alsoreported increased HNO3 coverages with increasingP(HNO3), albeit at P(HNO3) values substantially higherthan those presented here (5.0 � 10�7 to 3.0 � 10�6 hPa)[Hudson et al., 2002; Hynes et al., 2002]. These results

highlight the need for a comprehensive laboratory study ofHNO3 uptake on ice surfaces at P(HNO3) values below2.0 � 10�7 hPa that are typical of the subtropical UT.[21] The coverage of HNO3 on ice, in general, can be

modeled or predicted using the kinetics or thermodynamicsof the uptake process [Gao et al., 2003; Hudson et al.,2002]. Using laboratory measurements and a semiempiricalequilibrium surface coverage model, Hudson et al. [2002]have predicted HNO3 coverage on ice surfaces as a functionof temperature and P(HNO3). This multilayer Frenkel-Halsey-Hill (FHH) model was fitted to equilibrium HNO3

coverages observed on vapor-deposited ice films at temper-atures between 213 K and 219 K with a P(HNO3) of 1.1 �10�6 hPa. HNO3 surface coverages predicted by the FHHmodel are shown as a function of temperature in Figure 11a,together with the HNO3 coverages observed in cirrus cloudsduring CRYSTAL-FACE. The isobaric lines representingthe modeled coverages are colored on the same scale as theobserved coverages according to the values of P(HNO3)input to the model. Figure 11a indicates better agreementbetween the modeled and observed HNO3 coverages attemperatures higher than approximately 205 K, while atlower temperatures the modeled coverages increase tovalues far greater than those observed at comparable tem-peratures and P(HNO3) values. The high model coverages

Figure 9. (a) Volume-weighted mean diameter of cirruscloud particles versus SAD measured during the flights on11, 13, 19, and 21 July 2002. Only data at SADs greater than200 mm2 cm�3 are shown. Data are colored according toambient temperature. The black line represents the terminalfall velocity for ice particles in the upper troposphere,calculated according to Meier and Hendricks [2002].(b) IWC (represented as a gas-phase equivalent volumemixing ratio) versus SAD. At 130 hPa and 200 K, an IWCof 100 ppmv is equivalent to an ice water concentration of14 mg m�3. Other details same as Figure 9a.

Figure 10. HNO3 coverage versus temperature formeasurements made in cirrus clouds at SADs greater than200 mm2 cm�3 on 11, 13, 19, and 21 July 2002. Symbols arecolored according to P(HNO3). Triangles at temperaturesless than 200 K indicate measurements made underconditions in which NAT is stable. Mean values of HNO3

coverage are plotted as the mean of 5 K temperature binsfrom 195 to 220 K (black squares). Error bars represent thestandard deviation in each temperature bin. Negative valuesof HNO3 coverage are included in the calculation of themean values. The dashed line at 1.9 � 1013 molecules cm�2

represents the mean HNO3 coverage observed duringCRYSTAL-FACE. A complete HNO3 monolayer is formedwhen the coverage reaches 1.0 � 1015 molecules cm�2.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

9 of 14

D06302

below 205 K may result from the fact that the model wasfitted to laboratory data at temperatures above 213 K, andthe coverages presented here, therefore, are extrapolated tolower temperatures where the uncertainty in the modelincreases.[22] A number of studies have described the uptake of

HNO3 on ice surfaces using a Langmuir surface chemistrymodel [Tabazadeh et al., 1999;Hynes et al., 2002;Meier andHendricks, 2002]. The Langmuir isotherm predicts the frac-tional HNO3 surface coverage (q) according to equation (2),

q ¼K1=2eq � P HNO3ð Þ1=2

1þ K1=2eq � P HNO3ð Þ1=2

; ð2Þ

where Keq represents the equilibrium adsorption constant,given by the ratio of the rates of adsorption and desorption(ka/kd) [Laidler and Meiser, 1982]. The value of q is unitywhen the HNO3 surface coverage reaches a completemonolayer (1.0 � 1015 molecules cm�2). We note that thesurface density of HNO3 molecules when forming acomplete monolayer is somewhat uncertain, and the densityof 1.0 � 1015 molecules cm�2 stated here should beconsidered an upper limit [Hudson et al., 2002]. If the HNO3

surface density is lower than 1.0 � 1015 molecules cm�2

for a complete monolayer, the resulting fractional surfacecoverages will be higher than those stated here. Applica-tion of the dissociative form of the Langmuir isotherm issupported here by spectroscopic studies of HNO3 uptakeon thin ice films at 211 K, which indicate, by the presenceof H3O

+ and NO3� ions on the ice surface, that HNO3

dissociates upon adsorption [Zondlo et al., 1997]. Thetemperature-dependent equilibrium adsorption constant inequation (2) can be calculated according to equation (3)[Adamson and Gast, 1997],

Keq ¼100 � NA � s0 � t02p �M � R � Tð Þ1=2

e�DHadsc�R�T

� �hPa�1; ð3Þ

where NA is Avogadro’s number (6.02 � 1023 mol�1), s0is the area of one adsorption site (10�19 m2), t0 is thetime constant for adsorbate oscillation perpendicular tothe surface (10�13 s), M is the molecular weight ofHNO3 (0.063 kg mol�1), R is the ideal gas constant(8.314 J mol�1 K�1), T is temperature (in K), c is a unitconversion factor (2.39 � 10�4 kcal J�1) and DHads is theadsorption enthalpy of HNO3 on ice (in kcal mol�1).[23] Using equations (2) and (3), a Langmuir isotherm

was fitted to the CRYSTAL-FACE observations of fractionalsurface coverage and P(HNO3) (Figure 12). Using themedian temperature of 208 K for the observations shownin Figure 9, the best fit to the experimental data was achievedwith a DHads of�11.0 kcal mol�1, or 46.0 kJ mol�1 (red linein Figure 12). Note that the data are largely bound by DHads

values of �10.0 kcal mol�1 and �12.0 kcal mol�1. Alsoshown in Figure 12 are Langmuir isotherms at the sametemperature for DHads values of �14.2 kcal mol�1 and�12.9 kcal mol�1 reported by Tabazadeh et al. [1999] andHynes et al. [2002], respectively. Fractional coveragespredicted using these previously published values of DHads

far exceed the coverages observed during CRYSTAL-FACE,

Figure 11. HNO3 coverage versus temperature for measurements made in cirrus clouds at SADs greaterthan 200 mm2 cm�3 on 11, 13, 19, and 21 July 2002. Symbols are colored according to P(HNO3).Triangles at temperatures less than 200 K indicate measurements made under conditions in which NAT isstable. Data are represented as 10-s averages. (a) Lines are isobars representing HNO3 coveragescalculated by the FHH HNO3 uptake model [Hudson et al., 2002] and are colored according to the sametemperature scale as the observed coverages (see legend). (b) Same as Figure 11a, except the isobarsrepresent HNO3 coverages calculated by the Langmuir surface chemistry model.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

10 of 14

D06302

indicating these adsorption enthalpies are too high to accu-rately describe the observations presented here. Bartels-Rausch et al. [2002], using a chromatographic technique,have recently reported a DHads for HNO3 uptake on ice of�10.5 kcal mol�1 that is in good agreement with the value of�11.0 kcal mol�1 presented here. We caution that theeffective DHads reported here is empirically derived fromobservations in a dynamic system which may or maynot be in steady state, and this value, therefore, cannotbe considered a fundamental thermodynamic parameter.Nonetheless, the Langmuir formalism, using a DHads of�11.0 kcal mol�1, effectively describes the CRYSTAL-FACE observations of HNO3 uptake on cirrus cloud particlesto within a factor of 5 (Figure 12). It should be noted that thedata shown in Figure 12 span a temperature range form197 K to 218 K, and that the Langmuir isotherms were fittedat the median temperature of 208 K. Use of a singletemperature in fitting the isotherms is supported by theresults of Hynes et al. [2002], who reported a variation ofless than 2% between values of DHads derived from labora-tory measurements at 218 K and 228 K.[24] Having derived an effective value of DHads for

HNO3 adsorption on cirrus cloud particles in the UT,HNO3 coverages predicted by the Langmuir surface chem-istry model (as a function of temperature) can be comparedto the CRYSTAL-FACE observations (Figure 11b). As inFigure 11a, the isobaric lines are colored on the samescale as the observed coverages. The calculated coveragesshown in Figure 11b indicate that the model does notadequately describe the considerable variability in theobserved coverages at a given temperature and P(HNO3).

Nonetheless, when using the empirically derived DHads of�11.0 kcal mol�1, the Langmuir model is capable ofpredicting the observed coverages within a factor of 5 orbetter. The variability in the observed coverages, and theless than perfect agreement with the uptake models, can beexplained, in part, if the adsorbed HNO3 is not in equilib-rium with HNO3 in the gas phase. Previous field studieshave also shown HNO3 surface coverages to be highlyvariable throughout the temperature and P(HNO3) rangesobserved [Kondo et al., 2003; Ziereis et al., 2004].

4.2. HNO3 Partitioning in Cirrus Clouds

[25] The fraction of total HNO3 present on cirrus cloudparticles was observed to increase with SAD duringCRYSTAL-FACE (Figure 13). The mean value of HNO3

partitioned in the condensed phase at SADs greater than200 mm2 cm�3 was 16%. Up to 100% of the total HNO3 waspartitioned on ice particles during some cirrus cloud encoun-ters, at SADs between 350 and 4.2 � 104 mm2 cm�3 andtemperatures between 201 K and 213 K. Measurementsreported by Ziereis et al. [2004] in midlatitude cirrus cloudsreveal a similar relationship between condensed-phase NOy

partitioning and SAD, although maximum reported values ofFigure 12. Fractional HNO3 surface coverage (q) versusP(HNO3) measured during the flights of 11, 13, 19, and21 July 2002. Colored lines are isotherms fitted according toequations (2) and (3) (see text) with values of DHads shownin the legend and the median observed temperature of208 K. All data are represented as 10-s averages formeasurements made at SADs greater than 200 mm2 cm�3.

Figure 13. Fraction of total HNO3 in the condensed phaseversus SAD measured during the flights of 11, 13, 19, and21 July 2002. Data are represented as 10-s averages formeasurements made at SADs greater than 200 mm2 cm�3

and are colored according to temperature. Black squaresrepresent mean values of the 10-s data grouped intoquintiles. Vertical bars represent the standard deviationabout the mean value in each quintile and horizontal barsrepresent the upper and lower boundaries of each quintile.Values of condensed-phase partitioning greater than 100%occur when zero or near-zero HNO3jgas abundances aremeasured as negative values. The dashed line at 16%represents the mean value of HNO3 partitioned in thecondensed phase during CRYSTAL-FACE.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

11 of 14

D06302

condensed-phase NOy partitioned in cirrus clouds did notexceed 50% of the total observed NOy. Measurements ofNOy uptake in a mountain wave cirrus cloud reported byWeinheimer et al. [1998] indicate complete uptake of HNO3,provided that the ambient HNO3/NOy ratio in the cloud was0.1–0.2. Gas-phase HNO3 was not measured in either ofthese previous studies, making an accurate assessment of thefraction of HNO3 remaining in the gas phase after uptakedifficult.[26] Kramer et al. [2003] have recently studied the

partitioning of HNO3 in Arctic cirrus clouds, and havemodeled the role of HNO3 uptake by interstitial HNO3-H2SO4-H2O ternary solution aerosols in partitioning. Thisstudy concluded that some fraction of the total HNO3 inArctic cirrus clouds must remain in the gas phase, with theremainder partitioned predominately in interstitial aerosolsat temperatures less than 205 K when SADs are low, andon cirrus cloud particles at higher SADs. Measurements insubtropical cirrus clouds reported here, however, indicatethat up to 100% of the total HNO3 can be partitioned incirrus ice particles both at low temperatures and lowSADs. Furthermore, we see no evidence of significantuptake of HNO3 in ternary solution aerosols outside ofclouds in the subtropical UT during the flights consideredin this study. There is evidence, however, of HNO3 uptakeby ternary solution aerosols in the near absence of cirrusice particles on at least one other CRYSTAL-FACE flight(9 July 2002) (A. J. Weinheimer et al., manuscript inpreparation, 2004). We also note that HNO3 may be

contained in a ternary solution on the surface of the cirrusice particles.

4.3. HNO3 and HNO3//NOy in the Cloud-Free UpperTroposphere and Lower Stratosphere

[27] Measurements in the cloud-free subtropical UTduring CRYSTAL-FACE indicate that the gas-phaseHNO3/NOy ratio is highly variable, ranging from zero toapproximately 0.5 (colored symbols in Figure 14a). TheHNO3/NOy ratio is generally higher and also variable in thesubtropical lower stratosphere (LS), with values observedbetween 0.05 and 1 (black symbols in Figure 14a). HNO3 isexpected to be the predominate NOy species in the LS awayfrom the tropopause region [Neuman et al., 2001]. Thelower HNO3/NOy ratios (<0.3) observed in the UT areaffected by low observed values of HNO3, due to HNO3

removal by uptake and sedimentation by cloud particles incirrus processed air masses (yellow symbols in Figure 14b),or from elevated levels of NOy due to NO production fromlightning strikes (purple symbols in Figure 14b). Previouslymeasured values of the HNO3/NOy ratio in the midlatitudeUT over the continental United States were also highlyvariable and ranged from approximately 0.1 to 0.5 [Neumanet al., 2001]. The large range and variability of HNO3/NOy

ratios observed in the UT during CRYSTAL-FACE high-lights the value in measuring gas-phase HNO3 when assess-ing HNO3 uptake on cirrus cloud particles, over derivinggas-phase HNO3 from measured NOy and a constantassumed HNO3/NOy ratio.

Figure 14. (a) Vertical profile of the gas-phase HNO3/NOy ratio observed in cloud free air (SADs lessthan 20 mm2 cm�3) during the flights of 11, 13, 19, and 21 July 2002. Tropospheric measurements(according to the Microwave Temperature Profiler) are colored according to measured NO. Stratosphericmeasurements are shown in black. (b) Vertical profile of HNO3jgas. Other details same as Figure 14a.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

12 of 14

D06302

[28] A number of cirrus clouds observed in the subtrop-ical UT during CRYSTAL-FACE had VMDs between200 mm and 700 mm. Terminal fall velocities for cirrusice particles in this size range are 1 m s�1 to 10 m s�1

(Figure 9a) [Meier and Hendricks, 2002]. With 16% of thetotal HNO3 in cirrus clouds adsorbed on ice particles,the gravitational redistribution of a significant fraction ofthe total HNO3 in these clouds can therefore occur on atimescale of minutes to hours. Cirrus clouds with up to100% of the total HNO3 partitioned in the condensedphase have even greater potential to redistribute HNO3 inthe UT.

5. Conclusions and Implications

[29] A number of cirrus cloud encounters in the UT by theNASA WB-57F during CRYSTAL-FACE were accompa-nied by the observation of condensed-phase HNO3. Maxi-mum levels of condensed-phase HNO3 exceeded the gas-phase equivalent of 0.8 ppbv during some cirrus events. Amean HNO3 surface coverage of 1.9� 1013 molecules cm�2

was observed on the flights of 11, 13, 19, and 21 July 2002,with maximum surface coverages reaching as high as 1.4 �1014 molecules cm�2 during a few cirrus cloud encounters.Molecular coverages predicted using a Langmuir surfacechemistry model agree with the observed coverages towithin a factor of 5 or better when using an empiricallyderived DHads of�11.0 kcal mol�1. The mean percentage oftotal HNO3 condensed on cirrus cloud particles was 16%,with up to 100% of the HNO3 partitioned in the condensedphase in a number of cirrus clouds. The fraction of totalHNO3 in the condensed phase was found to increase stronglywith SAD. Based on the large diameters of cloud particlescontaining HNO3 observed during CRYSTAL-FACE, theredistribution of HNO3 in the UT will be very effective insome cloud systems. The interpretation of future observa-tions of HNO3 uptake on cirrus particles will be improved bya knowledge of the individual history of the air parcels inwhich the cirrus clouds are formed [Karcher, 2003] andlaboratory studies of HNO3 uptake at low gas-phase abun-dances (P(HNO3) < 2.0 � 10�7 hPa) and low temperatures(195 K–220 K).

[30] Acknowledgments. The authors wish to thank the air andground crews of the NASA WB-57F aircraft. Helpful discussions withJ. P. D. Abbatt, E. R. Lovejoy, and H. Ziereis are also appreciated. Thiswork was partially supported by the National Aeronautics and SpaceAdministration Upper Atmospheric Research Program and RadiationScience Program. Work done by MJM at the Jet Propulsion Laboratory,California Institute of Technology, was carried out under contract with theNational Aeronautics and Space Administration.

ReferencesAbbatt, J. P. D. (1997), Interaction of HNO3 with water-ice surfaces attemperatures of the free troposphere, Geophys. Res. Lett., 24(12),1479–1482.

Adamson, A. W., and A. P. Gast (1997), Physical Chemistry of Surfaces,6th Ed., John Wiley, New York.

Albritton, D. L., et al. (2001), Technical summary of the Working Group Ireport, in Climate Change 2001: The Scientific Basis. Contribution ofWorking Group I to the Third Assessment Report of the Intergovernmen-tal Panel on Climate Change, edited by J. T. Houghton et al., pp. 21–83,Cambridge Univ. Press, New York.

Arora, O. P., et al. (1999), Uptake of nitric acid by sub-micron-sized iceparticles, Geophys. Res. Lett., 26(24), 3621–3624.

Bartels-Rausch, T., et al. (2002), The adsorption enthalpy of nitrogenoxides on crystalline ice, Atmos. Chem. Phys., 2, 235–247.

Baumgardner, D., et al. (2001), The cloud, aerosol and precipitation spec-trometer: A new instrument for cloud investigations, Atmos. Res., 59–60,251–264.

Denning, R. F., S. L. Guidero, G. S. Parks, and B. L. Gary (1989), Instru-ment description of the Airborne Microwave Temperature Profiler,J. Geophys. Res., 94, 16,757–16,765.

Domine, F., and E. Thibert (1996), Mechanism of incorporation of tracegases in ice grown from the gas phase, Geophys. Res. Lett., 23(24),3627–3630.

Gao, R. S., et al. (2003), Evidence that ambient nitric acid increases relativehumidity in low-temperature cirrus clouds, Science, in press.

Gerber, H., Y. Takano, T. J. Garrett, and P. V. Hobbs (2000), Nephelometermeasurements of the asymmetry parameter, volume extinction coeffi-cient, and backscatter ratio in arctic clouds, J. Atmos. Sci., 57, 3021–3034.

Hanson, D., and K. Mauersberger (1988), Laboratory studies of the nitricacid trihydrate: Implications for the south polar stratosphere, Geophys.Res. Lett., 15(8), 855–858.

Hanson, D., and A. R. Ravishankara (1991), The reaction probabilities ofClONO2 and N2O5 on polar stratospheric cloud materials, J. Geophys.Res., 96(3), 5081–5090.

Heymsfield, A. J., and G. M. McFarquhar (2002), Mid-latitude and tropicalcirrus: Microphysical properties, in Cirrus, edited by D. K. Lynch et al.,pp. 78–101, Oxford Univ. Press, New York.

Hudson, P. K., et al. (2002), Uptake of nitric acid on ice at tropospherictemperatures: Implications for cirrus clouds, J. Phys. Chem. A, 106,9874–9882.

Hynes, R. G., M. A. Fernandez, and R. A. Cox (2002), Uptake of HNO3 onwater-ice and coadsorption of HNO3 and HCl in the temperature range210 – 235 K, J. Geophys. Res., 107(D24), 4797, doi:10.1029/2001JC001557.

Jaegle, L., et al. (1998), Sources of HOx and production of ozone in theupper troposphere over the United States, Geophys. Res. Lett., 25(10),1709–1712.

Jin, Y., W. B. Rossow, and D. P. Wylie (1996), Comparison of the climatol-ogies of high-level clouds from HIRS and ISCCP, J. Clim., 9, 2850–2879.

Karcher, B. (2002), Properties of subvisible cirrus clouds formed by homo-geneous freezing, Atmos. Chem. Phys., 2, 161–170.

Karcher, B. (2003), Simulating gas-aerosol-cirrus interactions: Process-oriented microphysical model and applications, Atmos. Chem. Phys., 3,1645–1664.

Kondo, Y., et al. (2003), Uptake of reactive nitrogen on cirrus cloud par-ticles in the upper troposphere and lowermost stratosphere, Geophys. Res.Lett., 30(4), 1154, doi:10.1029/2002GL016539.

Kramer, M., et al. (2003), Nitric acid partitioning in cirrus clouds: Asynopsis based on field, laboratory and model studies, Atmos. Chem.Phys. Discuss., 3, 413–443.

Lacis, A. A., D. J. Wuebbles, and J. A. Logan (1990), Radiative forcing ofclimate by changes in the vertical distribution of ozone, J. Geophys. Res.,95(D7), 9971–9982.

Laidler, K. J., and J. H. Meiser (1982), Physical Chemistry, Benjamin-Cummings, Menlo Park, Calif.

Lawrence, M. G., and P. J. Crutzen (1998), The impact of cloud particlegravitational settling on soluble trace gas distributions, Tellus, Ser. B, 50,263–289.

Liao, X., W. B. Rossow, and D. Rind (1995), Comparison between SAGE IIand ISCCP high-level clouds: 1. Global and zonal mean cloud amounts,J. Geophys. Res., 100(D1), 1121–1135.

Liou, K.-N. (1986), Influence of cirrus clouds on weather and climateprocesses: A global perspective, Mon. Weather Rev., 114, 1167–1199.

Lynch, D. K. (2002), Cirrus history and definition, in Cirrus, edited byD. K. Lynch et al., pp. 3–10, Oxford Univ. Press, New York.

Meier, A., and J. Hendricks (2002), Model studies on the sensitivityof upper tropospheric chemistry to heterogeneous uptake of HNO3 oncirrus ice particles, J. Geophys. Res., 107(D23), 4696, doi:10.1029/2001JD000735.

Meilinger, S. K., et al. (1999), HNO3 partitioning in cirrus clouds, Geophys.Res. Lett., 26(14), 2207–2210.

Neuman, J. A., et al. (1999), Study of inlet materials for sampling atmo-spheric nitric acid, Environ. Sci. Technol., 33, 1133–1136.

Neuman, J. A., et al. (2000), A fast-response chemical ionization massspectrometer for in situ measurements of HNO3 in the upper troposphereand lower stratosphere, Rev. Sci. Instrum., 71(10), 3886–3892.

Neuman, J. A., et al. (2001), In situ measurements of HNO3, NOy, NO, andO3 in the lower stratosphere and upper troposphere, Atmos. Environ., 35,5789–5797.

Northway, M. J., et al. (2002), An analysis of large HNO3-containingparticles sampled in the Arctic stratosphere during the winter of 1999/2000, J. Geophys. Res., 107(D20), 8298, doi:10.1029/2001JD001079.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

13 of 14

D06302

Sassen, K. (2002), Cirrus clouds: A modern perspective, in Cirrus, editedby D. K. Lynch et al., pp. 11–40, Oxford Univ. Press, New York.

Scott, S. G., T. P. Bui, K. R. Chan, and S. W. Bowen (1990), The meteoro-logical measurement system on the NASA ER-2 aircraft, J. Atmos.Oceanic Technol., 7, 525–540.

Sommerfeld, R. A., C. A. Knight, and S. K. Laird (1998), Diffusion ofHNO3 in ice, Geophys. Res. Lett., 25(6), 935–938.

Tabazadeh, A., O. B. Toon, and E. J. Jensen (1999), A surface chemistrymodel for nonreactive trace gas adsorption on ice: Implications for nitricacid scavenging by cirrus, Geophys. Res. Lett., 26(14), 2211–2214.

Vincent, J. H., et al. (1986), On the aspiration characteristics of large-diameter thin-walled aerosol sampling probes at yaw orientations withrespect to the wind, J. Aerosol Sci., 17(2), 211–224.

Wang, P.-H., et al. (1996), A 6-year climatology of cloud occurrencefrequency from Stratospheric Aerosol and Gas Experiment II observa-tions (1985–1990), J. Geophys. Res., 101(D23), 29,407–29,429.

Weinheimer, A. J., et al. (1998), Uptake of NOy on wave-cloud ice parti-cles, Geophys. Res. Lett., 25(10), 1725–1728.

Weinstock, E. M., et al. (1994), New fast-response photofragment fluores-cence hygrometer for use on the NASA ER-2 and the Perseus remotelypiloted aircraft, Rev. Sci. Instrum., 65(11), 3544–3554.

Wylie, D. P., and W. P. Menzel (1999), Eight years of high cloud statisticsusing HIRS, J. Clim., 12, 170–184.

Ziereis, H., et al. (2004), Uptake of reactive nitrogen on cirrus cloud particlesduring INCA, Geophys. Res. Lett., 31, doi:10.1029/2003GL018794,in press.

Zondlo, M. A., S. B. Barone, and M. A. Tolbert (1997), Uptake of HNO3

on ice under upper tropospheric conditions, Geophys. Res. Lett., 24(11),1391–1394.

�����������������������D. Baumgardner, Universidad Nacional Autonoma de Mexico, Centro de

Ciencias de la Atmosfera, Ciudad Universitaria, 04510 Mexico DF,Mexico.T. P. Bui, NASA Ames Research Center, Moffett Field, CA 94035, USA.S. Dhaniyala, Department of Mechanical and Aeronautical Engineering,

Clarkson University, Potsdam, NY 13699, USA.D. W. Fahey, R. S. Gao, P. K. Hudson, T. P. Marcy, P. J. Popp, and T. L.

Thompson, Aeronomy Laboratory, National Oceanic and AtmosphericAdministration, Boulder, CO 80305, USA. (ppopp@al.noaa.gov)T. J. Garrett, Department of Meteorology, University of Utah, Salt Lake

City, UT 84112, USA.B. Karcher, DLR, Institut fur Physik der Atmosphare, D-82234 Wessling,

Germany.D. J. Knapp, D. D. Montzka, B. A. Ridley, and A. J. Weinheimer,

Atmospheric Chemistry Division, National Center for AtmosphericResearch, Boulder, CO 80307, USA.M. J. Mahoney, Jet Propulsion Laboratory, California Institute of

Technology, Pasadena, CA 91109, USA.J. V. Pittman, D. S. Sayres, J. B. Smith, and E. M. Weinstock,

Atmospheric Research Project, Harvard University, Cambridge, MA 02138,USA.

D06302 POPP ET AL.: NITRIC ACID UPTAKE ON CIRRUS PARTICLES

14 of 14

D06302